H. BLOOM AND N. J. DOULL
620
Vol. 60
moles of solvent are expanded by 41 cc. and the value is 34.4 e.u. The difference, 1.4 e.u., must (liquid) It then dissolved in it, the contribution of represent mainly the shortcoming of the simple the expansion to the entropy would be 6.7 cal. F-H equation with respect to this system. We deg.-'. This agrees remarkably well with the postpone a more detailed consideration of the value 6.9 calculated earlier by Hildebrand and matter pending completion of the study of the broScott.2 The sum of the entropy of fusion, expan- mine-f-heptane system. sion and dilution to x2 = 1.M X lo-' is thus 8.0 The support of the Atomic Energy Commission 6.7 18.3 = 33.0 e.u. The experimental is gratefully acknowledged.
+
+
TRANSPORT NUMBER MEASUREMENTS I N PURE FUSED SALTS BY H. BLOOM AND N. J. DOULL Department of Chemistry, Auckland University College, Auckland, New Zealand Received November 8 , 1066
I n all previous attem t s to determine transport numbers in molten salts, the ravitational flow of melt in the reverse direction to that produce! by the flow of electricity, has introduced large errors wtich have never been completely overcome by the insertion of narrow tubes or sintered disks between the anode and cathode compartments. I n the present work, gravitational flow has been eliminated completely by the use of an accurately horizontal transport cell. The transport numbers of the chloride ion in molten lead chloride and cadmium chloride have been determined by measuring the movement of electrolyte in a capillary tube during the pasmge of a known quantity of electricity. The values of the measured transport number are PbC12, t- = 0.393 f 0.01 (527-529"),t- = 0.382 & 0.01 (602408"); CdCh, t- = 0.340 i 0.007 (602-608'). These values indicate that previous assumptions that these salts are predominantly anion conductors were incorrect. No evidence has been found ,for the presence of autocomplexes in these salts.
Comparison of the electrical conductivities of molten salts has led to the assumption that conduction in the molten alkali halides is predominantly cationic but the conduction process in the alkaline earths together with those of lead and cadmium is mainly anionic.'V2 To verify such assumptions determinations of transport numbers in pure fused salts are necessary. For molten salts such measurements have not, in the past, yielded reliable results because, as Duke and Laity3s4have pointed out, the transfer of electrolyte by the passage of current in a Hittorf type apparatus builds up a hydrostatic pressure which tends to nullify the electrolytic movement of the melt. Previous workers have attempted to counteract this gravitational flow by inserting a deterrent to such flow between the anode and cathode compartments. These devices, such as sintered glass disks or narrow capillary tubes, will slow down but do not .prevent the gravitational movement of the melt. Lorenz and Fausti6 attempted to determine the transport number of chloride ion in fused mixtures of lead chloride and potassium chloride. Two small porous cells immersed in the fused salt mixture formed the anode and cathode compartments, respectively. The measurements were carried out a t about 800" and were inconclusive. Wirthse used cells separated into three compartments by sintered disks for the investigation of lead chloride and added lead chloride containing radioactive lead (Th B) to the center compartment so as to permit radiochemical analysis. Temperature fluctuations and gravitational flow in the reverse (1) E. Heymann and H. Bloom, Nature, 166, 479 (1945). Roy. SOC.(London), 188A,392 (2) H. Bloom and E. Heymrtnn, PTOC. (1947). (3) F. R. Duke and R. W. Laity, J . A m . Chem. SOC.,76, 4090 (1954). (4) F. R. Duke and R. W. Laity, THISJOURNAL, 69, 549 (1955). ( 5 ) R. Lorenz and G. Fausti, 2. EEektrochem., 10, 630 (1904). (6) G. Wirtlis, ibid., 48, 480 (1937).
direction made the transport measurements inconclusive. Baimakov and Samusenko' also used a Hittorf type apparatus for investigation of lead chloride but without success. Karpachev and Pal'guev* used a similar method to investigate lead chloride but obtained results of poor reproducibility. Duke and Laityso4 modified the usual Hittorf method. Their Pyrex glass transport cell consisted of two compartments separated by a sintered glass disk and joined above the disk by a capillary. Two lead pools a t the bottom of the compartments formed the electrodes and the apparatus was filled so as to leaye an air bubble in the capillary. To carry out a run, the air bubble was displaced by adding a weighed amount of powdered lead chloride to one compartment and electrolysis was carried out until the bubble returned to its original position, the quantity of electricity being measured. I n this method the error due to gravitational flow of electrolyte in the opposite direction to that of electrolysis was not eliminated. This will be discussed below. There also have been attempts t o investigate other salt~.~J~ The present research project was planned so as to eliminate errors due to gravitational flow of the melt in the direction opposite to that of the electrolysis. Experimental Materials.-These were all of analytical reagent purity. Lead Chloride was prepared by precipitation from analytical reagent lead nitrate solution by analytical reagent hydrochloric acid. The product was filtered and evaporated to dryness several times with hydrochloric acid to remove any traces of nitrate. (7) Yu. V. Baimakov and 8. P. Samusenko, Trans. Leningiad I n d . Inat. (1938), No. 1, Sect. Met. No. 1, 3-26. (8) 8. Karpaahev and 8. Pal'guev, Zhuv. Fie. K h i m . , 98, 942 (1949). (9) K. E.Schwarz, 2.Elektrochem., 47, 144 (1941). (10) P. M. Aziz and F. E. W. Wetmore. Canadian J . Chem., 80, 779 (1952),
May, 1956
TRANSPORT NUMBER MEASUREMENTS IN PURE FUSED SALTS
Cadmium Chloride was prepared by direct chlorination of molten cadmium, purity 99.99%, which was supplied by Electrolytic Zinc Co. of Australia. Apparatus.-The transport cell was constructed of transgarent silica. Pyrex was first tried but was abandoned ecause of deformation above 560’ 553” is given as the annealing temperature of Pyrex”). he cell consisted of two compartments (internal diameter 6 mm., length of each 1 cm.) separated by a sintered silica disk. Disks of various porosities from 2 (medium) to 4 (fine) were used in different runs without any systematic variation of results. Each compartment terminated in a capillary of length 10 cm. and bore 0.5 to 1.0 mm. These ca illaries were calibrated and checked for uniformity by weigiing with mercury inside. Tungsten wires of diameter 0.01 inch were introduced into the ca illaries to carry the electrolysis current to the molten metayelectrodes. A transparent silica tube wound with nichrome V wire (8 turns per inch) formed the heating element. This element was mounted in an insulating case containing slits through which its contents could be viewed. By focusing a light source and cathetometer telescope on the boundary between molten salt and molten metal in the ca illary, its movement during electrolysis could be measurex Temperature was measured by means of a chromel-alumel thermocouple in contact with the transport cell and a Leeds and Northrup type K2 potentiometer. Temperature was controlled by a variable voltage transformer and did not vary more than 0.5” during a run. The current source WM carefully stabilized,12 the current being measured by determination of potential dro across a standard resistance in the circuit, by means of t i e potentiometer. The furnace was mounted so that i t could be rotated in the vertical plane perpendicular to its axis. This allowed the transport cell to be filled in a vertical position, then rotated to a horizontal position to carr out the run. Experimental Procedure.-$o fill the transport cell, the empty cell with tungsten wires removed was attached at one end to a vacuum system and the cell waB heated in the vertical furnace. Lead chloride melt was drawn into the bulk of the cell and then a small quantity of lead to form one of the electrodes. When the furnace was made horizontal the tungsten wires were introduced into the capillariesone into the molten lead and the other into the lead chloride. The latter was made the cathode so that after a preliminary electrolysis this wire also became surrounded by molten lead. The current was then stopped and any movement of the boundary between lead and lead chloride noted. Levelling adjustments of the furnace were made until there was no gravitational movement of the boundary in 1 hour. T o carry out a determination of transport number, current was passed at 10 ma. and the rate of movement of the anolyte boundary measured over a time interval of 1-2 hours.
cr
Temp. range, “C.
PbCla CdCli
527-529 602-608 602-608
621
Transport no.
t = 0.393 f.0.01 t- =: 0.382 f 0.01 t0.340 i 0.007
The determinations were not sufficiently accurate to decide whether there is a real variation of transport number with change of temperature. It is clear however that the cations in molten PbCl2 and CdC12 carry relatively more of the current than the anions. This is in agreement with the postulate by Bloom and Heymann2 that the smaller ion will be expected to carry most of the current. These authors however further stated that PbC12and CdC12may be expected to be anion conductors because they are anion conductors in the solid state and the conductivity of molten lead and cadmium halides changes considerably with change of anion. It can also be seen that autocomplex formation, involving the presence of lead and cadmium in the form of large complex anions, e.g., PbClB4-, is ruled out by the results obtained. If autocomplexes were present in appreciable concentration the transport number of the cation would be very low or even negative. Considering the relative magnitude of the t+ values of PbCL and CdCL (where t+ = 1 - t-) it can be seen that t+ PbC‘2 = 0.94 which is approxit+ CdClz mately equal to Apbc12/Acdc12 at corresponding temperatures (10% in degrees absolute above the melting point2)-A refers to equivalent conductivity. This would be anticipated as we are comparing salts with common anion, hence the ratio of their conductances should be equal to the ratio of the cation mobilities. The results for anion transport number in lead chloride do not agree with those obtained by Karpachev and Pal’guev or Duke and Laity, who obtained the value of t- = 0.75. The latter authors criticized the work of Karpachev and Pal’guev even though their results agreed with those of the Russian workers. Duke and Laity’s Results and Discussion method does not overcome the problem of gravitaAccording to Duke and LaityS14the transport tional flow through the sintered disk in the opposite number of the chloride ion is equal to the fraction direction to that of electrolysis. These authors of an equivalent of salt transferred from catholyte reduced the hydrostatic head causing this flow to anolyte during the passage of one faraday of to very small values but in doing so they measured a electricity. The transport number of the anion is correspondingly small movement of electrolyte. The net error due to gravitational flow therefore thus given by the equation remains the same. To overcome the problem of gravitational flow Duke and Laity recommend the use of “ultrafine” disks-coarse disks did not give them any meaningful results. I n the present where work there is no gravitational flow as the apparatus V = increase in vol. of molten salt in the anode compart- was carefully levelled before each run, the porous ment disk being present merely to enable the detection q = no. of coulombs of electricity passed d = density of salt of the movement of electrolyte relative to the M = molecular weight of salt bulk of the melt. Other criticisms which apply to the experiments Similar experiments were conducted to determine the anion transport number in cadmium chloride of Duke and Laity are as follows: (1) Pyrex glass a t 5 6 5 O , Le., slightly above its annealing temperawith molten cadmium electrodes. ture (553’) is likely to be slightly plastic and the (11) C. J. Phillips, “Glass-The Miracle Maker,” Pitman, New relatively bulky apparatus may be subject to deYork, 1941. formation at this temperature. (2) The tempera(12) M. Spiro and H. N. Parton, Trans. Faraday Soc., 48, 263 ture was ‘(not too carefully controlled” which may (1952).
622
DAVID M. MASONAND STEPHEN P. VANGO
Vol. 60
introduce errors due to thermal expansion both of ages varying from 0.2 to 6 volts were applied across the melt (if the apparatus is not perfectly sym- the transport cell without any systematic variation metrical) and the air bubble. (3) The electrolysis of the measured transport number. The authors are pleased to acknowledge assistcurrents used by Duke and Laity (up to 0.5 amp.) ance from the University of New Zealand Research cause considerable local heating of the disk. The present authors agree with Duke and Laity Fund for the purchase of the apparatus, and the that the disk does not introduce errors due to many helpful discussions with N. E. Richards and electroosmosis. I n the present investigation volt- J. Barton.
MEASUREMENT OF OPTICAL ABSORBANCY OF TiCl,(g) IN THE ULTRAVIOLET REGION' BY DAVID M. MASON^ AND STEPHEN P. VANGO Jet Propulsion Laboratory, California Institute of Technology, Pasadena, Calif. Received November 4 , 2966
__
The optical absorbancy of gaseous TiCL was measured a t 25" a t wave lengths from 240 to 360 mp. A maximum in absorbancy occurs a t 280 mp where the molar absorbancy index was found t o be 7.29 X 109 liter/mole cm. The partial TiClr(g) TiClt(s) was measured by opticalabsorbancy and found pressure of TiC14(g) in the equilibrium 2TiCla(s) t o be about 1 mm. at 475'. By successively removing TiClr and measuring the optical absorbancy of the system, solid solution of Tic12 in TiCls was shown t o be non-existent in this equilibrium.
I. Introduction Few data are available on the absorption spec-
+
trum of gaseous TiC14 in the ultraviolet range.a Such data were obtained to supplement kinetic and thermodynamic studies involving TiCla, particularly the equilibrium of disproportionation of TiCla(s). I n measurement of this equilibrium using the Knudsen effusion meth~d,~J'initially high partial pressures of TiC14(g) were obtained which leveled out a t a lower value as the measurement progressed. This behavior suggested solid solution of TiClz(s) in TiC13(s) causing initially low activity of TiClz(s) and therefore high Tic14 pressure. This behavior was studied in an isochoric system using absorbancy to measure equilibrium pressures of Tic&. The absorbancy of TiClr(g) was first measured as a function of wave length for use with these other physiochemical measurements. 11. Experimental The absorbance measurements were made with a Beckman Model DU spectrophotometer using quartz cells either 1 cm. or 5 cm. in depth. A vertical quartz stem, which was bent over and sealed to a bulb, was attached to the cell. The bulb was kept in a constant-temperature liquid bath outside the spectrophotometer. This bulb was partially filled with TiC14(1)of 99.999% purity provided by the National Bureau of Standards. The T i c 4 was frozen by immersing the bulb in liquid nitrogen, and the glass system was then evacuated and sealed. The cell temperature was maintained to within f0.1" by controlling a special air bath surrounding the spectrophotometer cell holder. Most of the measurements were made with the cell a t around 2 5 O , although several measurements were made at 50' to check possible deviations from Beer's law. The concentration of TiCL(g) was varied by varying the temperature of the liquid TiCL in the temperature range 0 to 8". The molar absorbancy index was calculated from the measured absorbance together with vapor pressure data for TiCh.6
WAVELENGTH ( m p ) .
Fig. 1.,-Optical absorbancy of TiCldg). (1)
This paper presents the results of one phaae of research carried
out at the Jet Propulsion Laboratory, Caliiornia Institute of Technology, sponsored by the Department of the Army, Ordnance Corps, and the Department of the Navy, Office of Naval Research, under Contract No. DA-04-495-ORD 18.
(2) Department of Chemistry and Chemical Engineering, Stanford University, Stanford, California. (3) Y. Hukumoto, Sci. Repts. TbhokuImp. Uniu.. [I]!22,868(1933). (4) M. Farber and A. J. Darnell, to be published. (5) G . E. MacWood and B. S. Sanderson 111, "Disproportionation Equilibrium of Titanium Trichloride," Technical Report No. 4, Contract NR 037-024. Columbus, Ohio, Ohio State University Research Foundation, 1955. (6) K. K. Kelley, "The Free Energies of Vaporization and Vapor Pressures of Inorganic Substance," Bureau of Mines, Bulletin No. 383, 1935.
